AUTO-UPDATED

Anthropic warns of ‘catastrophic’ AI risks in its own IPO filing

Anthropic is preparing for a historic $2 trillion IPO while disclosing significant financial losses, heavy infrastructure spending, and potential existential risks associated with its advanced artificial intelligence models.

Key Points

  • Anthropic reported a $42 billion net loss in 2025 despite generating $4.6 billion in revenue.
  • The company plans to invest $518 billion in cloud and computing infrastructure to support its AI development.
  • Nearly 25% of the company's 2025 revenue originated from only two clients.
  • The IPO prospectus dedicates 80 pages to potential AI risks, including models exhibiting self-preserving behaviors and information manipulation.
  • A "Founder LLC" structure will grant the seven cofounders 50.1% of total voting power to maintain control post-IPO.

Why it Matters

This filing highlights the extreme capital intensity and inherent safety risks associated with scaling frontier AI models. Investors must weigh the potential for massive market growth against the company's precarious revenue concentration and the existential liabilities inherent in its technology.
The Verge Published by Jess Weatherbed
Read original